openfoam/tutorials/verificationAndValidation/atmosphericModels/HargreavesWright_2007

#------------------------------------------------------------------------------

Overview

    "By setting appropriate profiles for wind velocity and the turbulence
    quantities at the inlet, it is often assumed that the boundary layer will
    be maintained up to the buildings or obstructions in the flow." (HW:p. 355).
    However, it was quantified by (HW:p. 355) that "even in the absence of
    obstructions, ..., the velocity and turbulence profiles decay along the
    fetch" (HW:p. 355). It was shown by (HW:p. 355) that a set of modifications
    were required to maintain a neutral atmospheric boundary layer throughout
    an empty and long computational domain of a RANS computation.

    Aim:

        Verification of the following boundary conditions in terms of the
        maintenance of inlet quantities downstream within a RANS computation:

        - atmBoundaryLayerInletVelocity
        - atmBoundaryLayerInletK
        - atmBoundaryLayerInletEpsilon
        - atmBoundaryLayerInletOmega

    Benchmark (Physical phenomenon):

        The benchmark is an empty fetch computational domain, steady-state
        RANS simulation involving the following traits:

        - External flow
        - The surface layer portion of the neutral-stratified equilibrium
        atmospheric boundary layer (no Ekman layer)
        - Dry air
        - Homogeneous, smooth terrain
        - Spatiotemporal-invariant aerodynamic roughness length
        - No displacement height
        - Newtonian, single-phase, incompressible, non-reacting

    Benchmark scenario:

        - Computational domain: (HW:Fig. 1)
        - Benchmark dataset: (HW:Fig. 6) (Obtained by the WebPlotDigitizer-4.2
        (Rohatgi, 2019))

    Resources:

        Computational study (tag:HW):
            Hargreaves, D. M., & Wright, N. G. (2007).
            On the use of the k–ε model in commercial CFD software
            to model the neutral atmospheric boundary layer.
            Journal of wind engineering and
            industrial aerodynamics, 95(5), 355-369.
            DOI:10.1016/j.jweia.2006.08.002

        Wind profile (tag:RQP):
            Richards, P. J., Quinn, A. D., & Parker, S. (2002).
            A 6 m cube in an atmospheric boundary layer flow-Part 2.
            Computational solutions. Wind and structures, 5(2_3_4), 177-192.
            DOI:10.12989/was.2002.5.2_3_4.177

    Physical modelling:

        - The governing equations for:
            - Steady-state, Newtonian, single-phase, incompressible fluid flows,
            excluding any thermal chemical, electromagnetic and scalar
            interactions
        - Mathematical approach for the turbulence modelling:
            - Reynolds-averaged Navier-Stokes simulation (RANS)
        - Turbulence closure model:
            - kEpsilon and kOmegaSST linear eddy viscosity closure models
        - The sets of input (HW:Table 1):
            - Reference height, Zref = 6 [m]
            - Aerodynamic roughness height, z0 = 0.01 [m]
            - Displacement height, d = 0 [m]
            - Reference mean wind speed, Uref = 10 [m/s]

    Computational domain modelling:

        - Rectangular prism
        - (x1, x2, x3)
            = (5000, 100, 500) [m]
            = (streamwise, spanwise, ground-normal) directions

    Computational domain discretisation:

        - Spatial resolution:
            - (x1, x2, x3) = (500, 5, 50) [cells]
            - Refer to the `system/blockMeshDict` for the grading details
        - Temporal resolution: Steady state

    Equation discretisation:

        - Spatial derivatives and variables:
            - Convection: Second order
            - Others: Second order with various limiters

        - Temporal derivatives and variables: First order

    Numerical boundary/initial conditions:

        - Refer to `0.orig`

    Pressure-velocity coupling algorithm:

        - SIMPLEC

    Linear solvers:

        - Refer to `system/fvSolution`

    Initialisation and sampling:

        - No initialisation/averaging
        - Sampling at the end of the simulation via `system/sampleDict`
        - Refer to `system/controlDict` for further details


#------------------------------------------------------------------------------